All-in-one AI framework & toolkit for Claude Code & Cursor

Trellis is an AI framework and toolkit designed to enhance consistency and productivity for developers using AI coding assistants like Claude Code and Cursor. It addresses the challenge of maintaining AI output quality and project context by automating the injection of best practices and enabling persistent, collaborative AI workflows.

How It Works

Trellis employs an auto-injection mechanism to embed predefined specifications (workflow guides, coding standards, patterns) directly into AI conversations, ensuring deterministic and consistent code generation. Its core advantage lies in a layered architecture that uses context compression, loading only relevant specs for the current task, unlike monolithic AI configuration files. Features like parallel sessions in isolated worktrees, team-wide spec synchronization, and session persistence via project journaling further streamline development and collaboration.

Quick Start & Requirements

• Primary install / run command: npm install -g @mindfoldhq/trellis@latest followed by trellis init -u your-name in your project directory.
• Non-default prerequisites and dependencies: Node.js and npm.
• The -u flag sets up a personal workspace at .trellis/workspace/your-name/.

Highlighted Details

• Enforced Standards: Trellis enforces specifications via hook injection, guaranteeing adherence unlike optional AI skills.
• Layered Context: Utilizes a layered architecture with context compression for efficient, task-specific spec loading.
• Parallel Workflows: Enables concurrent development by spawning multiple AI agents in isolated worktrees.
• Persistent Memory: AI retains project context across sessions through automatic journaling and git integration.

Maintenance & Community

Developed by Mindfold. Community engagement is facilitated via a Discord server and bug reporting/feature requests through GitHub Issues. The roadmap indicates ongoing development with planned enhancements for code review, skill packs, and broader tool integration.

Licensing & Compatibility

The project is released under the FSL (Free Software License). Specific FSL variant details are not provided, which may impact commercial use or redistribution terms.

Limitations & Caveats

The specific terms of the FSL license require clarification for commercial applications. The project is actively developed, with key features like advanced code review and broader IDE support still on the roadmap, suggesting it may not yet be fully mature or feature-complete for all use cases.